arXiv subjects

Jiaqi Wang

Publications and source records attributed to Jiaqi Wang.

At least 235 records · Page 13Linked to original sources

arXiv2025

Foundation Models in Federated Learning: Assessing Backdoor Vulnerabilities

Federated Learning (FL), a privacy-preserving machine learning framework, faces significant data-related challenges. For example, the lack of suitable public datasets leads to ineffective information exchange, especially in heterogeneous environments with uneven data distribution. Foundation Models (FMs) offer a promising solution by generating synthetic datasets that mimic client data distributions, aiding model initialization and knowledge sharing among clients. However, the interaction between FMs and FL introduces new attack vectors that remain largely unexplored. This work therefore assesses the backdoor vulnerabilities exploiting FMs, where attackers exploit safety issues in FMs and poison synthetic datasets to compromise the entire system. Unlike traditional attacks, these new threats are characterized by their one-time, external nature, requiring minimal involvement in FL training. Given these uniqueness, current FL defense strategies provide limited robustness against this novel attack approach. Extensive experiments across image and text domains reveal the high susceptibility of FL to these novel threats, emphasizing the urgent need for enhanced security measures in FL in the era of FMs.

cs.CR↗

arXiv2025

IDArb: Intrinsic Decomposition for Arbitrary Number of Input Views and Illuminations

Capturing geometric and material information from images remains a fundamental challenge in computer vision and graphics. Traditional optimization-based methods often require hours of computational time to reconstruct geometry, material properties, and environmental lighting from dense multi-view inputs, while still struggling with inherent ambiguities between lighting and material. On the other hand, learning-based approaches leverage rich material priors from existing 3D object datasets but face challenges with maintaining multi-view consistency. In this paper, we introduce IDArb, a diffusion-based model designed to perform intrinsic decomposition on an arbitrary number of images under varying illuminations. Our method achieves accurate and multi-view consistent estimation on surface normals and material properties. This is made possible through a novel cross-view, cross-domain attention module and an illumination-augmented, view-adaptive training strategy. Additionally, we introduce ARB-Objaverse, a new dataset that provides large-scale multi-view intrinsic data and renderings under diverse lighting conditions, supporting robust training. Extensive experiments demonstrate that IDArb outperforms state-of-the-art methods both qualitatively and quantitatively. Moreover, our approach facilitates a range of downstream tasks, including single-image relighting, photometric stereo, and 3D reconstruction, highlighting its broad applications in realistic 3D content creation.

cs.CV↗

arXiv2025

Multi-label classification for multi-temporal, multi-spatial coral reef condition monitoring using vision foundation model with adapter learning

Coral reef ecosystems provide essential ecosystem services, but face significant threats from climate change and human activities. Although advances in deep learning have enabled automatic classification of coral reef conditions, conventional deep models struggle to achieve high performance when processing complex underwater ecological images. Vision foundation models, known for their high accuracy and cross-domain generalizability, offer promising solutions. However, fine-tuning these models requires substantial computational resources and results in high carbon emissions. To address these challenges, adapter learning methods such as Low-Rank Adaptation (LoRA) have emerged as a solution. This study introduces an approach integrating the DINOv2 vision foundation model with the LoRA fine-tuning method. The approach leverages multi-temporal field images collected through underwater surveys at 15 dive sites at Koh Tao, Thailand, with all images labeled according to universal standards used in citizen science-based conservation programs. The experimental results demonstrate that the DINOv2-LoRA model achieved superior accuracy, with a match ratio of 64.77%, compared to 60.34% achieved by the best conventional model. Furthermore, incorporating LoRA reduced the trainable parameters from 1,100M to 5.91M. Transfer learning experiments conducted under different temporal and spatial settings highlight the exceptional generalizability of DINOv2-LoRA across different seasons and sites. This study is the first to explore the efficient adaptation of foundation models for multi-label classification of coral reef conditions under multi-temporal and multi-spatial settings. The proposed method advances the classification of coral reef conditions and provides a tool for monitoring, conserving, and managing coral reef ecosystems.

cs.CV↗

arXiv2025

OVO-Bench: How Far is Your Video-LLMs from Real-World Online Video Understanding?

Temporal Awareness, the ability to reason dynamically based on the timestamp when a question is raised, is the key distinction between offline and online video LLMs. Unlike offline models, which rely on complete videos for static, post hoc analysis, online models process video streams incrementally and dynamically adapt their responses based on the timestamp at which the question is posed. Despite its significance, temporal awareness has not been adequately evaluated in existing benchmarks. To fill this gap, we present OVO-Bench (Online-VideO-Benchmark), a novel video benchmark that emphasizes the importance of timestamps for advanced online video understanding capability benchmarking. OVO-Bench evaluates the ability of video LLMs to reason and respond to events occurring at specific timestamps under three distinct scenarios: (1) Backward tracing: trace back to past events to answer the question. (2) Real-time understanding: understand and respond to events as they unfold at the current timestamp. (3) Forward active responding: delay the response until sufficient future information becomes available to answer the question accurately. OVO-Bench comprises 12 tasks, featuring 644 unique videos and approximately human-curated 2,800 fine-grained meta-annotations with precise timestamps. We combine automated generation pipelines with human curation. With these high-quality samples, we further developed an evaluation pipeline to systematically query video LLMs along the video timeline. Evaluations of nine Video-LLMs reveal that, despite advancements on traditional benchmarks, current models struggle with online video understanding, showing a significant gap compared to human agents. We hope OVO-Bench will drive progress in video LLMs and inspire future research in online video reasoning. Our benchmark and code can be accessed at https://github.com/JoeLeelyf/OVO-Bench.

cs.CV↗

arXiv2025

VaLiD: Mitigating the Hallucination of Large Vision Language Models by Visual Layer Fusion Contrastive Decoding

Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities in multimodal task reasoning. However, they often generate responses that appear plausible yet do not accurately reflect the visual content, a phenomenon known as hallucination. Recent approaches have introduced training-free methods to mitigate hallucinations by adjusting the decoding strategy during the inference stage, typically attributing hallucinations to the language model itself. Our analysis, however, reveals that distortions in the visual encoding process significantly affect the model's reasoning capabilities. Specifically, earlier visual layers may retain key features but gradually distort as the information propagates toward the output layer. Building on these insights, we propose a novel hallucination-mitigation method from the visual encoding perspective: \textbf{V}isu\textbf{a}l \textbf{L}ayer Fus\textbf{i}on Contrastive \textbf{D}ecoding (\textbf{VaLiD}). This method utilizes uncertainty to guide the visual layer selection, correcting distortions in the visual encoding process and thereby enhancing the reliability of the generated content. Experimental results demonstrate the effectiveness of VaLiD in mitigating hallucinations across various benchmarks, achieving state-of-the-art performance when compared to baseline methods. Codes are available at \href{https://github.com/RicardoLuL/VaLiD_LVLMs_hallucinations}{Github}.

cs.CV↗

arXiv2025

Light-A-Video: Training-free Video Relighting via Progressive Light Fusion

Recent advancements in image relighting models, driven by large-scale datasets and pre-trained diffusion models, have enabled the imposition of consistent lighting. However, video relighting still lags, primarily due to the excessive training costs and the scarcity of diverse, high-quality video relighting datasets. A simple application of image relighting models on a frame-by-frame basis leads to several issues: lighting source inconsistency and relighted appearance inconsistency, resulting in flickers in the generated videos. In this work, we propose Light-A-Video, a training-free approach to achieve temporally smooth video relighting. Adapted from image relighting models, Light-A-Video introduces two key techniques to enhance lighting consistency. First, we design a Consistent Light Attention (CLA) module, which enhances cross-frame interactions within the self-attention layers of the image relight model to stabilize the generation of the background lighting source. Second, leveraging the physical principle of light transport independence, we apply linear blending between the source video's appearance and the relighted appearance, using a Progressive Light Fusion (PLF) strategy to ensure smooth temporal transitions in illumination. Experiments show that Light-A-Video improves the temporal consistency of relighted video while maintaining the relighted image quality, ensuring coherent lighting transitions across frames. Project page: https://bujiazi.github.io/light-a-video.github.io/.

cs.CV↗

arXiv2025

GPT4Scene: Understand 3D Scenes from Videos with Vision-Language Models

In recent years, 2D Vision-Language Models (VLMs) have made significant strides in image-text understanding tasks. However, their performance in 3D spatial comprehension, which is critical for embodied intelligence, remains limited. Recent advances have leveraged 3D point clouds and multi-view images as inputs, yielding promising results. However, we propose exploring a purely vision-based solution inspired by human perception, which merely relies on visual cues for 3D spatial understanding. This paper empirically investigates the limitations of VLMs in 3D spatial knowledge, revealing that their primary shortcoming lies in the lack of global-local correspondence between the scene and individual frames. To address this, we introduce GPT4Scene, a novel visual prompting paradigm in VLM training and inference that helps build the global-local relationship, significantly improving the 3D spatial understanding of indoor scenes. Specifically, GPT4Scene constructs a Bird's Eye View (BEV) image from the video and marks consistent object IDs across both frames and the BEV image. The model then inputs the concatenated BEV image and video frames with markers. In zero-shot evaluations, GPT4Scene improves performance over closed-source VLMs like GPT-4o. Additionally, we prepare a processed video dataset consisting of 165K text annotation to fine-tune open-source VLMs, achieving state-of-the-art performance on all 3D understanding tasks. Surprisingly, after training with the GPT4Scene paradigm, VLMs consistently improve during inference, even without object marker prompting and BEV image as explicit correspondence. It demonstrates that the proposed paradigm helps VLMs develop an intrinsic ability to understand 3D scenes, which paves the way for a seamless approach to extending pre-trained VLMs for 3D scene understanding.

cs.CV↗

arXiv2025

Visual-RFT: Visual Reinforcement Fine-Tuning

Reinforcement Fine-Tuning (RFT) in Large Reasoning Models like OpenAI o1 learns from feedback on its answers, which is especially useful in applications when fine-tuning data is scarce. Recent open-source work like DeepSeek-R1 demonstrates that reinforcement learning with verifiable reward is one key direction in reproducing o1. While the R1-style model has demonstrated success in language models, its application in multi-modal domains remains under-explored. This work introduces Visual Reinforcement Fine-Tuning (Visual-RFT), which further extends the application areas of RFT on visual tasks. Specifically, Visual-RFT first uses Large Vision-Language Models (LVLMs) to generate multiple responses containing reasoning tokens and final answers for each input, and then uses our proposed visual perception verifiable reward functions to update the model via the policy optimization algorithm such as Group Relative Policy Optimization (GRPO). We design different verifiable reward functions for different perception tasks, such as the Intersection over Union (IoU) reward for object detection. Experimental results on fine-grained image classification, few-shot object detection, reasoning grounding, as well as open-vocabulary object detection benchmarks show the competitive performance and advanced generalization ability of Visual-RFT compared with Supervised Fine-tuning (SFT). For example, Visual-RFT improves accuracy by $24.3\%$ over the baseline in one-shot fine-grained image classification with around 100 samples. In few-shot object detection, Visual-RFT also exceeds the baseline by $21.9$ on COCO's two-shot setting and $15.4$ on LVIS. Our Visual-RFT represents a paradigm shift in fine-tuning LVLMs, offering a data-efficient, reward-driven approach that enhances reasoning and adaptability for domain-specific tasks.

cs.CV↗

arXiv2025

Enhancing Monocular 3D Scene Completion with Diffusion Model

3D scene reconstruction is essential for applications in virtual reality, robotics, and autonomous driving, enabling machines to understand and interact with complex environments. Traditional 3D Gaussian Splatting techniques rely on images captured from multiple viewpoints to achieve optimal performance, but this dependence limits their use in scenarios where only a single image is available. In this work, we introduce FlashDreamer, a novel approach for reconstructing a complete 3D scene from a single image, significantly reducing the need for multi-view inputs. Our approach leverages a pre-trained vision-language model to generate descriptive prompts for the scene, guiding a diffusion model to produce images from various perspectives, which are then fused to form a cohesive 3D reconstruction. Extensive experiments show that our method effectively and robustly expands single-image inputs into a comprehensive 3D scene, extending monocular 3D reconstruction capabilities without further training. Our code is available https://github.com/CharlieSong1999/FlashDreamer/tree/main.

cs.GR↗

arXiv2025

OmniAlign-V: Towards Enhanced Alignment of MLLMs with Human Preference

Recent advancements in open-source multi-modal large language models (MLLMs) have primarily focused on enhancing foundational capabilities, leaving a significant gap in human preference alignment. This paper introduces OmniAlign-V, a comprehensive dataset of 200K high-quality training samples featuring diverse images, complex questions, and varied response formats to improve MLLMs' alignment with human preferences. We also present MM-AlignBench, a human-annotated benchmark specifically designed to evaluate MLLMs' alignment with human values. Experimental results show that finetuning MLLMs with OmniAlign-V, using Supervised Fine-Tuning (SFT) or Direct Preference Optimization (DPO), significantly enhances human preference alignment while maintaining or enhancing performance on standard VQA benchmarks, preserving their fundamental capabilities. Our datasets, benchmark, code and checkpoints have been released at https://github.com/PhoenixZ810/OmniAlign-V.

cs.CV↗

arXiv2025

ByTheWay: Boost Your Text-to-Video Generation Model to Higher Quality in a Training-free Way

The text-to-video (T2V) generation models, offering convenient visual creation, have recently garnered increasing attention. Despite their substantial potential, the generated videos may present artifacts, including structural implausibility, temporal inconsistency, and a lack of motion, often resulting in near-static video. In this work, we have identified a correlation between the disparity of temporal attention maps across different blocks and the occurrence of temporal inconsistencies. Additionally, we have observed that the energy contained within the temporal attention maps is directly related to the magnitude of motion amplitude in the generated videos. Based on these observations, we present ByTheWay, a training-free method to improve the quality of text-to-video generation without introducing additional parameters, augmenting memory or sampling time. Specifically, ByTheWay is composed of two principal components: 1) Temporal Self-Guidance improves the structural plausibility and temporal consistency of generated videos by reducing the disparity between the temporal attention maps across various decoder blocks. 2) Fourier-based Motion Enhancement enhances the magnitude and richness of motion by amplifying the energy of the map. Extensive experiments demonstrate that ByTheWay significantly improves the quality of text-to-video generation with negligible additional cost.

cs.CV↗

arXiv2025

PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction

In large vision-language models (LVLMs), images serve as inputs that carry a wealth of information. As the idiom "A picture is worth a thousand words" implies, representing a single image in current LVLMs can require hundreds or even thousands of tokens. This results in significant computational costs, which grow quadratically as input image resolution increases, thereby severely impacting the efficiency of both training and inference. Previous approaches have attempted to reduce the number of image tokens either before or within the early layers of LVLMs. However, these strategies inevitably result in the loss of crucial image information, ultimately diminishing model performance. To address this challenge, we conduct an empirical study revealing that all visual tokens are necessary for LVLMs in the shallow layers, and token redundancy progressively increases in the deeper layers of the model. To this end, we propose PyramidDrop, a visual redundancy reduction strategy for LVLMs to boost their efficiency in both training and inference with neglectable performance loss. Specifically, we partition the LVLM into several stages and drop part of the image tokens at the end of each stage with a pre-defined ratio, creating pyramid-like visual tokens across model layers. The dropping is based on a lightweight similarity calculation with a negligible time overhead. Extensive experiments demonstrate that PyramidDrop can achieve a 40% training time and 55% inference FLOPs acceleration of LLaVA-NeXT with comparable performance. Besides, the PyramidDrop could also serve as a plug-and-play strategy for inference acceleration without training, with better performance and lower inference cost than counterparts. Code is available at https://github.com/Cooperx521/PyramidDrop.

cs.CV↗

arXiv2025

BoostStep: Boosting mathematical capability of Large Language Models via improved single-step reasoning

Large language models (LLMs) have demonstrated impressive ability in solving complex mathematical problems with multi-step reasoning and can be further enhanced with well-designed in-context learning (ICL) examples. However, this potential is often constrained by two major challenges in ICL: granularity mismatch and irrelevant information. We observe that while LLMs excel at decomposing mathematical problems, they often struggle with reasoning errors in fine-grained steps. Moreover, ICL examples retrieved at the question level may omit critical steps or even mislead the model with irrelevant details. To address this issue, we propose BoostStep, a method that enhances reasoning accuracy through step-aligned ICL, a novel mechanism that carefully aligns retrieved reference steps with the corresponding reasoning steps. Additionally, BoostStep incorporates an effective "first-try" strategy to deliver exemplars highly relevant to the current state of reasoning. BoostStep is a flexible and powerful method that integrates seamlessly with chain-of-thought (CoT) and tree search algorithms, refining both candidate selection and decision-making. Empirical results show that BoostStep improves GPT-4o's CoT performance by 4.6% across mathematical benchmarks, significantly surpassing traditional few-shot learning's 1.2%. Moreover, it can achieve an additional 7.5\% gain combined with tree search. Surprisingly, it enhances state-of-the-art LLMs to solve challenging math problems using simpler examples. It improves DeepSeek-R1-671B's performance on AIME by 2.2%, leveraging simple examples only from the MATH dataset.

cs.CL↗

arXiv2025

RelightVid: Temporal-Consistent Diffusion Model for Video Relighting

Diffusion models have demonstrated remarkable success in image generation and editing, with recent advancements enabling albedo-preserving image relighting. However, applying these models to video relighting remains challenging due to the lack of paired video relighting datasets and the high demands for output fidelity and temporal consistency, further complicated by the inherent randomness of diffusion models. To address these challenges, we introduce RelightVid, a flexible framework for video relighting that can accept background video, text prompts, or environment maps as relighting conditions. Trained on in-the-wild videos with carefully designed illumination augmentations and rendered videos under extreme dynamic lighting, RelightVid achieves arbitrary video relighting with high temporal consistency without intrinsic decomposition while preserving the illumination priors of its image backbone.

cs.CV↗

arXiv2025

OneLLM: One Framework to Align All Modalities with Language

Multimodal large language models (MLLMs) have gained significant attention due to their strong multimodal understanding capability. However, existing works rely heavily on modality-specific encoders, which usually differ in architecture and are limited to common modalities. In this paper, we present OneLLM, an MLLM that aligns eight modalities to language using a unified framework. We achieve this through a unified multimodal encoder and a progressive multimodal alignment pipeline. In detail, we first train an image projection module to connect a vision encoder with LLM. Then, we build a universal projection module (UPM) by mixing multiple image projection modules and dynamic routing. Finally, we progressively align more modalities to LLM with the UPM. To fully leverage the potential of OneLLM in following instructions, we also curated a comprehensive multimodal instruction dataset, including 2M items from image, audio, video, point cloud, depth/normal map, IMU and fMRI brain activity. OneLLM is evaluated on 25 diverse benchmarks, encompassing tasks such as multimodal captioning, question answering and reasoning, where it delivers excellent performance. Code, data, model and online demo are available at https://github.com/csuhan/OneLLM

cs.CV↗

arXiv2025

Discrete Wavelet Transform-Based Capsule Network for Hyperspectral Image Classification

Hyperspectral image (HSI) classification is a crucial technique for remote sensing to build large-scale earth monitoring systems. HSI contains much more information than traditional visual images for identifying the categories of land covers. One recent feasible solution for HSI is to leverage CapsNets for capturing spectral-spatial information. However, these methods require high computational requirements due to the full connection architecture between stacked capsule layers. To solve this problem, a DWT-CapsNet is proposed to identify partial but important connections in CapsNet for a effective and efficient HSI classification. Specifically, we integrate a tailored attention mechanism into a Discrete Wavelet Transform (DWT)-based downsampling layer, alleviating the information loss problem of conventional downsampling operation in feature extractors. Moreover, we propose a novel multi-scale routing algorithm that prunes a large proportion of connections in CapsNet. A capsule pyramid fusion mechanism is designed to aggregate the spectral-spatial relationships in multiple levels of granularity, and then a self-attention mechanism is further conducted in a partially and locally connected architecture to emphasize the meaningful relationships. As shown in the experimental results, our method achieves state-of-the-art accuracy while keeping lower computational demand regarding running time, flops, and the number of parameters, rendering it an appealing choice for practical implementation in HSI classification.

cs.CV↗

arXiv2025

Dispider: Enabling Video LLMs with Active Real-Time Interaction via Disentangled Perception, Decision, and Reaction

Active Real-time interaction with video LLMs introduces a new paradigm for human-computer interaction, where the model not only understands user intent but also responds while continuously processing streaming video on the fly. Unlike offline video LLMs, which analyze the entire video before answering questions, active real-time interaction requires three capabilities: 1) Perception: real-time video monitoring and interaction capturing. 2) Decision: raising proactive interaction in proper situations, 3) Reaction: continuous interaction with users. However, inherent conflicts exist among the desired capabilities. The Decision and Reaction require a contrary Perception scale and grain, and the autoregressive decoding blocks the real-time Perception and Decision during the Reaction. To unify the conflicted capabilities within a harmonious system, we present Dispider, a system that disentangles Perception, Decision, and Reaction. Dispider features a lightweight proactive streaming video processing module that tracks the video stream and identifies optimal moments for interaction. Once the interaction is triggered, an asynchronous interaction module provides detailed responses, while the processing module continues to monitor the video in the meantime. Our disentangled and asynchronous design ensures timely, contextually accurate, and computationally efficient responses, making Dispider ideal for active real-time interaction for long-duration video streams. Experiments show that Dispider not only maintains strong performance in conventional video QA tasks, but also significantly surpasses previous online models in streaming scenario responses, thereby validating the effectiveness of our architecture. The code and model are released at \url{https://github.com/Mark12Ding/Dispider}.

cs.CV↗

arXiv2024

Asymmetrical Reciprocity-based Federated Learning for Resolving Disparities in Medical Diagnosis

Geographic health disparities pose a pressing global challenge, particularly in underserved regions of low- and middle-income nations. Addressing this issue requires a collaborative approach to enhance healthcare quality, leveraging support from medically more developed areas. Federated learning emerges as a promising tool for this purpose. However, the scarcity of medical data and limited computation resources in underserved regions make collaborative training of powerful machine learning models challenging. Furthermore, there exists an asymmetrical reciprocity between underserved and developed regions. To overcome these challenges, we propose a novel cross-silo federated learning framework, named FedHelp, aimed at alleviating geographic health disparities and fortifying the diagnostic capabilities of underserved regions. Specifically, FedHelp leverages foundational model knowledge via one-time API access to guide the learning process of underserved small clients, addressing the challenge of insufficient data. Additionally, we introduce a novel asymmetric dual knowledge distillation module to manage the issue of asymmetric reciprocity, facilitating the exchange of necessary knowledge between developed large clients and underserved small clients. We validate the effectiveness and utility of FedHelp through extensive experiments on both medical image classification and segmentation tasks. The experimental results demonstrate significant performance improvement compared to state-of-the-art baselines, particularly benefiting clients in underserved regions.

cs.LG↗